No abstract

Infections are among the leading causes of death in patients with inflammatory bowel disease (IBD), likely reflecting sepsis-associated mortality. However, a recent single-centered study demonstrated lower mortality among septic patients with IBD, as compared to the general population and it was hypothesized that the immune dysregulation in IBD may have a protective effect in response to infection. The association of IBD with short-term mortality in sepsis has not been adequately characterized in large populations. Methods: The Texas Public Use Data File was used to identify hospitalizations aged ≥18 years with a diagnosis of sepsis during 2014-2017. Sepsis was identified using "explicit" ICD-9 and -10 codes for severe sepsis (995.92, R65.20) and for septic shock (785.52, R65.21). IBD was identified using ICD-9 and -10 codes for Crohn's disease [CD] (555.x, K50.x) and ulcerative colitis [UC] (556.x, K51.x). Mixed-effects multivariable logistic regression modeling was used to examine the association of IBD with short-term mortality (defined as in-hospital mortality or discharge to hospice) among patients with sepsis. Similar approach was used to model the prognostic impact of CD and UC and for sensitivity analyses of hospitalizations with ICU admission and those with septic shock.Results: There were 283,025 sepsis hospitalizations during the study period, of which 3,105 (1.1%) had IBD [CD 52%, UC 48%]. Compared to those without IBD, hospitalizations with IBD were younger (age ≥65 years, 46.3% vs 56.8%), with lower mean (SD) Deyo comorbidity index (2.03 [0.79] vs 2.24 [0.77]) [p<0.0001 for both comparisons], but had similar mean (SD) number of organ failures (2.45 [0.60] vs 2.46 [0.59]; p=0.6494). The short-term mortality among sepsis hospitalizations with and without IBD was 26.8% (CD 22.1%, UC 31.8%) and 31.3%, respectively. Short-term mortality among IBD hospitalizations remained lower on adjusted analyses (adjusted odds ratio [aOR] 0.83 [95% CI 0.76-0.91]). The findings on sensitivity analyses were consistent with the primary model. However, on subgroup analyses, the risk of short-term mortality was lower only in CD (aOR 0.71 [(95% CI 0.62-0.82]), but not in UC (aOR 0.96 [95% CI 0.84-1.09]).Conclusions: IBD was associated with better short-term outcome in sepsis. However, this "protective" association was driven exclusively by septic patients with CD, while presence of UC did not impact short-term mortality. Further studies are needed to elucidate the differences in immune response to infection between CD and UC, which may inform identification of therapeutic targets for sepsis care in the general population.